AI / ML Engineer
Agentic Systems & LLM Orchestration
Posted May 2026
We are looking for an AI/ML Engineer to join our core engineering team and work at the intersection of large language models, agentic system design, and enterprise software integration.
This is not a research role. You will be deploying production AI systems that handle real business operations for clients across financial services, SaaS, healthcare, and logistics. The problems you solve will be in production by the time the sprint closes.
You will work directly with our senior engineers and technical leads on the full arc of AI system development: from scoping and architecture through to deployment, observability, and continuous improvement. There are no hand-off walls here.
- Design and implement LLM orchestration pipelines using LangChain, LlamaIndex, and CrewAI — from single-chain inference to complex multi-agent workflows
- Build and deploy Retrieval-Augmented Generation (RAG) systems that ground model outputs in private, domain-specific client data with hallucination rates maintained below production thresholds
- Engineer autonomous AI agents capable of multi-step reasoning, tool use, and cross-system action — integrated with client CRMs, ERPs, document systems, and APIs
- Develop and maintain vector database infrastructure (Pinecone, Weaviate, pgvector) supporting semantic search and context retrieval at enterprise scale
- Implement model evaluation and observability frameworks — tracking accuracy, latency, token efficiency, and business-outcome metrics across deployed systems
- Collaborate with frontend and backend engineers to integrate AI capabilities into client-facing products, ensuring the intelligence layer is performant, secure, and seamlessly embedded
- Contribute to prompt engineering and fine-tuning workflows — iterating on system prompts, few-shot examples, and domain-specific model variants
- Participate in client discovery and architecture sessions — communicating technical decisions to non-technical stakeholders clearly and confidently
Solid Python engineering fundamentals
You write clean, maintainable, production-grade Python. You understand async patterns, API design, and how to structure a codebase that other engineers can work in confidently.
Hands-on LLM & agentic framework experience
You have built real things with LangChain, LlamaIndex, CrewAI, AutoGen, or comparable orchestration frameworks. You understand their internals well enough to debug them when they misbehave — which they will.
RAG architecture depth
You have designed and deployed RAG systems, understand the tradeoffs between chunking strategies, embedding models, and retrieval approaches, and have opinions about when RAG is and is not the right solution.
API integration fluency
You are comfortable integrating LLM systems with external APIs and enterprise data sources. REST, GraphQL, webhooks — you have worked across them.
Production mindset
You think about latency, token costs, failure modes, and monitoring from the start — not as an afterthought when something breaks at 2 AM.
Valued but not required
- Experience with model fine-tuning (LoRA, QLoRA, full fine-tune) on open-source models
- Familiarity with cloud infrastructure (AWS, GCP, Azure) and containerised deployments (Docker, Kubernetes)
- Frontend exposure — enough to understand how the AI layer surfaces to users
- Experience in a client-facing or consultancy environment
- Contributions to open-source AI/ML projects
You are intellectually curious in a way that translates into code. When a new model or framework ships, your first instinct is to understand how it actually works — not just what it claims to do.
You are comfortable with ambiguity. Enterprise AI problems rarely arrive with clean specifications. You can reason through an unclear brief, identify the constraints that matter, and make a defensible call on the right technical approach.
You take quality seriously — not as a process to follow, but as a standard you hold yourself to. You review your own code before asking anyone else to review it. You document decisions. You test edge cases.
You are a direct communicator. You can explain a complex architectural decision to a client who has never heard of a vector database — and do so without condescension or unnecessary jargon.
Competitive Compensation
Market-rate salary with performance review at 6 months. Details discussed in first conversation.
Hybrid Working
3 days in our Mumbai office, 2 days from wherever you work best.
Learning Budget
Dedicated time and budget for courses, certifications, and conferences.
Growth Trajectory
A small team means your scope expands with your ability — not with headcount cycles.
We respond to every application within 5 business days. No recruiter intermediaries. You'll hear from our team directly.